Long Notice? Take It Slow. 🦥 Building NoticeNap with Open-Weight AI

Description and use

NoticeNap turns lengthy PDF notices into a structured summary and action checklist, cutting manual search time.

Problem

Applicants struggle to locate deadlines, eligibility, fees, and instructions scattered across long PDF notices, wasting time searching manually.

Outcome

NoticeNap transforms a complex PDF notice into a concise, structured summary and actionable checklist, reducing manual search time and improving workflow efficiency.

Workflow

• Upload PDF
• Extract text page‑by‑page with PyMuPDF
• Send extracted text to local Gemma 2 via Ollama
• Generate structured output (deadlines, eligibility, fees, etc.)
• Validate output with Pydantic, marking missing info as "Not found" or "Needs verification"
• Create action checklist and export to Markdown/iCalendar
• Allow user to review source references and track progress

Evidence

NoticeNap is a local‑first AI assistant designed to make reading long, information‑heavy PDF notices less tedious.

Source

NoticeNap combines PDF text extraction, a Python backend, and an open‑weight language model served locally through Ollama.

Source

The backend uses Pydantic to validate structured data returned through the analysis workflow.

Source

NoticeNap organizes findings into a checklist, supports tracking task progress, and provides Markdown and calendar export functionality where implemented.

Source